Thermal and air quality effects on the performance of schoolwork by children David P. Wyon Technical University of Denmark Swegon Air Academy Leiden, NDL 07.04.2010 Field intervention experiments This lecture reports the findings of a recent series of field intervention experiments that was carried out in schools in Denmark and Sweden on behalf of ASHRAE DTU was awarded the research contract in open competition with bidders from USA, Europe, Australia and the Far East The Final Report was accepted in 2009
Final Report of ASHRAE 1257 The effects of temperature, outdoor air supply rate and airborne particles on children in school classrooms David P. Wyon and Pawel Wargocki Technical University of Denmark www.ie.dtu.dk Main RP-1257 Objective To determine whether improving classroom air quality and ensuring classrooms do not become warm can improve the performance of schoolwork by children Window opening behavior was passively recorded during the experiments
Methods Indoor air quality was modified by increasing the outdoor air supply rate from a mechanical ventilation system Air temperature was reduced in hot weather by operating split cooling units Windows could be opened as usual Window opening behavior was recorded as all windows closed or 11 or more open Door opening was also recorded Selected school in Denmark School at Rungsted DTU
Schools Elementary school with no IEQ problems South-facing facades Mechanically ventilated Energy conservation measures in place Plan of the selected classrooms Selected classrooms
Selected classrooms View of classroom
Ventilation system Split cooling Electrostatic air cleaners Split cooling was installed and either operated or idled in summer
Silent electrostatic air cleaners were installed and operated or idled 4th to 6th grade 10-12 12-year old ~300 pupils Pupils
Physical measurements Continuous measurements (with pupils): CO 2, Air T, RH, window opening state recorders Effective outdoor air supply rate in L/sp was estimated from CO 2 rate of increase each time children entered the classroom. Note that this includes air entering through windows as well as supply air Measurements of perceptions and symptoms The children marked Visual-Analogue scales at the end of the week Intensity of symptoms and their perceptions of the classroom environment was indicated How is the classroom right now? Too cold The air is It is draughty humid The air is poor Too dark Too noisy How do I feel right now? Nose is blocked Throat is dry Lips are dry Skin is dry I am very hungry I slept badly last I feel very night tired I slept too little last I do not feel night like I working have a headache Too warm The air is still It is dry The air is fresh Too much light Too dark I can breathe freely Throat is not dry Lips are not dry Skin is not dry I am full I slept well last night I am not at all tired I slept long last night I feel like working today I do not have a headache
Measurements of performance Tasks appropriate to children children s age, developed in consultation with class teachers 4 languagelanguage-based: Acoustic proofproof-reading Reading and comprehension Logical reasoning ProofProof-reading 4 numerical: Subtraction Multiplication Number comparison Addition Tasks performed during mathematics and language lessons Sensory assessments of air quality by visiting adult panel ACCEPTABILITY SCALE Clearly acceptable Just acceptable Just not acceptable Clearly not acceptable After classes ended
Permission was obtained from: Parents School Headmaster School Board Teachers Local Authority Danish Ethics Review Board Swedish Ethics Review Board Design of experiments Interventions always improved conditions Interventions were for 1 week at a time in balanced order of presentation Cross-over design (one classroom with existing conditions and one with improved conditions) The ventilation system was modified to provide more air to one classroom at a time Split cooling installed and run in one classroom at a time
2x2 design Temperature Current Cool Ventilation rate current high Estimated effective L/s p 10 8 L/s per person 6 4 2 AC No AC 0 Low L/s High L/s
Mean classroom CO2 ppm 1200 1000 Mean CO2 ppm 800 600 400 200 AC No AC 0 Low L/s High L/s Peak classroom CO2 ppm Peak CO2 ppm 1400 1200 1000 800 600 400 200 0 Low L/s High L/s AC No AC
Teaching environment and routines No restrictions on normal daily activities No changes in class schedules Doors and windows could be opened No contact between researchers and children Reduced classroom temperature 2 independent experiments in August/September 2004 and 2005: 1st experiment: temperature reduced from 23.5 to 20 o C at two ventilation rates 2nd experiment: temperature reduced from 25 to 21 o C with no mechanical ventilation Other parameters of classroom IEQ were unchanged
Typical time-course of classroom temperature in the two classrooms 30 28 26 T [ C ] 24 22 20 18 07:30 09:00 10:30 12:00 13:30 15:00 16:30 Time Children s s perception of temperature Too warm (P<0.001) Too cold Estimated PMV Low temp PMV=0 High temp PMV=0.7
Performance of schoolwork as a function of classroom temperature 1,4 Performance 1,2 1 0,8 R 2 = 0.68 0,6 18 20 22 24 26 Temperature 1 o C lower temperature ~3.5% higher performance Earlier experiments by DP Wyon in Sweden in 1967 Raising classroom temperature from 20 C to 30 C C reduced most types of schoolwork performance by up to 30% For 40 years this was widely believed to be an overestimate The present results confirm the size of the thermal effect first reported in 1967: +10 C would lead to 35% less schoolwork
Increased outdoor air supply rate Three independent field intervention experiments with 100% outdoor air supplied through new filters: 1. 3.4 was increased to 9.5 L/sp 2. 3.0 was increased to 6.5 L/sp 3. 5.0 was increased to 9.5 L/sp (with and without cooling 2200 2000 1800 Time-course of classroom CO 2 concentration, example ppm 1600 1400 1200 1000 800 600 400 15:59 15:30 15:02 14:33 14:04 13:35 13:06 12:38 12:09 11:40 11:11 10:42 10:14 09:45 09:16 08:47 08:18 07:50
Children s s perception of IAQ Fresh air (P<0.001) Poor air Low vent High vent Sensory assessments of air quality made by an untrained panel after children had left the classroom % (P<0.001) 40 Dissatisfied 30 20 10 0 Low vent High vent
Effect of increasing ventilation rate on performance: increase in speed % 45 35 25 15 5 Experiment 1 Experiment 2 Experiment 3 * * * * * * * * * -5 Subtraction Multiplication Number comparison Addition Logical reasoning Reading and comprehension Performance of schoolwork as a function of classroom ventilation 1,4 Performance 1,2 1 0,8 0,6 R 2 = 0.59 0 2 4 6 8 10 12 Ventilation (L/sp) Doubling ventilation rate ~14.5% higher performance
Classroom air quality matters Windows are often not opened Outdoor air supply rates are often low CO2 levels are often above 1000 ppm Increasing the air supply rate from 2.5 to 5 to 10 L/sp would improve the performance of schoolwork by 29% Poor air quality is a major disadvantage for children, and especially for slow workers CO2 in 1663 Scandinavian classrooms October 2009
Overall result from 1663 Scandinavian clasrooms CO2 in Danish classrooms Random selection of 100 classrooms Structured representative sample CO2 recorded throughout school hours 2 weeks in November 2009 Outdoor temperature was 0-150 C Proportion with mean CO2 > 1000 ppm : 66%
Results Condition of supply air filter Two studies in January 2005 No consistent effects on perceptions, symptoms or perceived air quality when a used filter was replaced with a new one Installing a new filter had no effect on children s s performance % 120 110 Speed (P<0.92) % 120 110 % errors (P<0.37) Performance 100 90 100 90 80 80 70 Used filter New filter 70 Used filter New filter
Interpretation Results differ from office findings because: Very little dust was retained in the used supply air filter Supply air was 100% outdoor air, so all the dust originated outdoors Operation of electrostatic air cleaners 2 independent experiments in January and March/April 2005 in 1 Danish and 4 Swedish schools Concentration of particles in classrooms was considerably reduced when electrostatic air cleaners were in operation
Number of particles per cm3 6000 5000 4000 Placebo units Air cleaners 3000 2000 1000 0 >0.75 >1.0 >2.0 >3.5 >5.0 >7.5 >10 >15 Concentration of particles per cm3 (with/without electrostatic deposition) Reduction greater in the 2 poorly ventilated classrooms 6000 5000 4000 3000 2000 1000 0
Operating electrostatic air cleaners had no effects on performance Performance % 120 110 100 90 80 Speed (P<0.59) % 120 110 100 90 80 % errors (P<0.36) 70 Placebo units Air cleaners 70 Placebo units Air cleaners Conclusions on filters and airborne particles Electrostatic air cleaners reduced the concentration of airborne particles but had no effects on schoolwork Installing a new filter in the outdoor air supply flow had no effects on schoolwork However, reducing airborne particle concentration may reduce the long-term effects of IAQ on health
Main 1257-RP conclusions Reducing even slightly warm classroom temperatures eliminated thermal discomfort and improved children s performance Increasing outdoor air supply rate improved classroom air quality and children s performance Schoolwork was performed faster, with no increase in errors, in both cases Implications Removing airborne particles had no effect, so the IAQ effects were due to gas-phase pollutants in classroom air Replacing supply air filters had no effect but increasing the outdoor air supply rate did, so pollutants must originate indoors The active pollutants could be bioeffluents or emissions from materials, or both
IAQ effects on adults Emissions: : performance of office work was reduced when office materials were present behind a screen in a clean room Bioeffluents: : performance of office work was reduced when the outdoor air supply was reduced in a clean room Conclusion: both classroom emissions and bioeffluents are likely to affect the performance of schoolwork by children Window opening Throughout the DK experiments, state loggers fitted to windows and doors recorded when they were open Teachers and children were able to open windows and doors as usual This is the first intervention experiment on window-opening opening behaviour in response to raised temperature or decreased outdoor air supply rate in school classrooms
Summertime classroom T in C (NB: Different classes with natural/mechanical ventilation) 30 25 20 No AC AC 15 Off Low High Mechanical ventilation Observations Split cooling reduced air T by only 1.7 C C at the highest ventilation rate (T was reduced by 5.4 C C at the lower rate) With natural ventilation, cooling was not very effective, reducing T by only 3.3 C (NB: this could be due to poor mixing, as T was recorded at 2.2m height)
Outdoor air supply rate (L/sp) (NB: Different classes with natural/mechanical ventilation) 10 5 No AC AC 0 Off Low High Mechanical ventilation Hours/week with windows open (NB: Different classes with natural/mechanical ventilation) 15 10 5 No AC AC 0 Off Low High Mechanical ventilation
Observations Windows were opened ed much more when it was warm in the classroom Even at 10 L/sp, a 1.7 C C increase in T increased window opening by 75% With no cooling, opening windows provided only 3.7 L/sp Split cooling reduced window opening by 65%, reducing the outdoor air supply rate to only 2.7 L/sp Classroom T 25 (NB: different classes in winter & summer) C 20 Summer Winter 15 Low High Ventilation (with no AC)
Outdoor air supply rate (L/sp) 10 (NB: different classes in winter & summer) 8 6 4 Summer Winter 2 0 Low High Ventilation (with no cooling) Hours/week with windows open 10 (NB: different classes in winter & summer) 8 6 4 Summer Winter 2 0 Low High Ventilation (with no cooling)
Hours/week with windows open and with cross ventilation 10 (2 classes in winter, 2 other classes in summer) 8 6 4 Windows open Cross ventilation 2 0 Winter Summer (With mechanical ventilation set to low in both seasons) 15 Hours per week with windows open and with cross ventilation (Summer, same 2 classes in all 4 conditions) 10 5 Windows open Cross ventilation 0 T low (A/C) T high (No A/C) (With mechanical ventilation set to low throughout)
Observations Windows were open 4x as long in summer (when classroom T was 5 C 5 C warmer) This raised the total outdoor air supply rate from 6.5 to 9.5 L/sp Both windows and doors were used adaptively to achieve cross-ventilation Window opening was triggered by warmth but not by poor air quality Conclusions on window opening Windows were opened to reduce classroom temperatures, not to improve IAQ Natural ventilation did not ensure adequate ventilation,, even with cross-ventilation Split cooling, by eliminating the perceived need to open windows, decreased the air quality still further In this study, both slightly raised T and poor IAQ decreased children s s performance
Our thanks t are due to ASHRAE, STVF, Asthma and Allergifond and Formas for funding Sophie Irgens (mid-term project student) Bartlomiej Matysiak, Line Jark, Maria Schaub- Hansen, Kasper Lynge Jensen, Mateusz Komenda (M.Sc. students) Cristina Pirvu, Henry Willem (Ph.D. students) Pupils, teachers, schools, municipalities School Technical Services in Rungsted & Lund Further reading ASHRAE Journal, October 2006 (Summary of T, IAQ effects) ASHRAE HVAC&R Research J, March 2007 (2 papers: T, IAQ effects) ASHRAE HVAC&R Research J, May 2008 (e-filters) Indoor Air 2008, September 2008 (Observations of window opening)